United States

The healthcare industry has accumulated vast amounts of clinical data, much of which has traditionally been unstructured, including medical records, clinical data, patient communications, and visit notes. Clinician-patient conversations form a crucial part of medical records, with the resulting medical note serving as the ground truth for future interactions and treatment plans. Generating concise and accurate SOAP notes is critical for quality patient care and is especially challenging in specialty care, where relevance, clarity, and adherence to clinician preferences are paramount. These requirements make general-purpose LLMs unsuitable for producing high-quality specialty notes. While recent LLMs like GPT-4 and Sonnet 3.5 have shown promise, their high cost, size, latency, and privacy issues remain barriers for many healthcare providers.

We introduce SpecialtyScribe, a modular pipeline for generating specialty-specific medical notes. It features three components: an Information Extractor to capture relevant data, a Context Retriever to verify and augment content from transcripts, and a Note Writer to produce high quality notes. Our framework and in-house models outperform similarly sized open-source models by over 12% on ROUGE metrics.
Additionally, these models match top closed-source LLMs&#39; performance while being under 1% of their size. We specifically evaluate our framework for oncology, with the potential for adaptation to other specialties.

NAACL 2025

SpecialtyScribe: Enhancing SOAP note Scribing for Medical Specialties using LLM&#39;s

oncology

medical documentation

soap notes

## Welcome to 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics!
This year, the conference is in Mexico City. NAACL was actually already planned for Mexico City in 2021, but due to the pandemic the entire conference was moved online. This year, finally, we get to go! So it is my sincere pleasure to welcome you to Mexico City, whether in person or virtually. Having the conference in Mexico City is a good opportunity to emphasize that NAACL is our flagship conference for ACL members not only in North America but also in Central and South America, even though NAACL has been bearing “North” in its name. At this year’s conference, we have a theme to match, with a theme track on the Languages of Latin America to showcase the linguistic diversity of the region.
 
The opportunity to present at NAACL should not depend on a researcher’s travel budget, or their family status. This is why it is so important to make virtual participation at NAACL as good an experience as possible – but we want to also provide a good experience for in-person participants. As a community, we are still working out the best way to do that. This year at NAACL, we are trying out a big virtual poster session ahead of the conference, with the hope that this will make make for a lively and interactive experience. At the same time, we are reducing virtual oral presentations, which seem to be particularly tricky to make to work well. A big thanks to the NAACL program chairs and to Luciana Benotti for all their ideas and work to improve the virtual experience. And participants, virtual as well as in-person: Please let us know what worked for you and what didn’t, so we can continue to improve hybrid conferences. 
 
I have been lucky to work with many amazing people. Without their insight, dedication and patience, and without the many hours of work they put in, NAACL would not have been possible. A huge thank you to the program chairs Helena Gomez, Kevin Duh, and Steve Bethard – you are the best! 
 
Finally, I would like to thank all authors, invited speakers and panelists, area chairs and reviewers, the volunteers organizing and chairing sessions, and all attendees, in-person and virtual. Thank you for helping us make NAACL 2024 come to life. 
 
Welcome and hope you all enjoy the conference! 
Katrin Erk 
The University of Texas at Austin 
NAACL-2024 General Chair 
*You can read the full Welcome message in the [Conference Handbook (downloadable)](https://drive.google.com/file/d/1H1NvW0VASQjkSCYgw3mr-3l4yYDSxz6B/view?usp=sharing)*

To access the event page you need to register [**here.**](http://acl.swoogo.com/naacl2024) 
Your access to the event page is limited based on your registration type. If you registered for workshops only, you will gain access to full workshops content on the day of the workshops program.

Please register!

NAACL 2024

2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics

The healthcare industry has accumulated vast amounts of clinical data, much of which has traditionally been unstructured, including medical records, clinical data, patient communications, and visit notes. Clinician-patient conversations form a crucial part of medical records, with the resulting medical note serving as the ground truth for future interactions and treatment plans. Generating concise and accurate SOAP notes is critical for quality patient care and is especially challenging in specialty care, where relevance, clarity, and adherence to clinician preferences are paramount. These requirements make general-purpose LLMs unsuitable for producing high-quality specialty notes. While recent LLMs like GPT-4 and Sonnet 3.5 have shown promise, their high cost, size, latency, and privacy issues remain barriers for many healthcare providers.

We introduce SpecialtyScribe, a modular pipeline for generating specialty-specific medical notes. It features three components: an Information Extractor to capture relevant data, a Context Retriever to verify and augment content from transcripts, and a Note Writer to produce high quality notes. Our framework and in-house models outperform similarly sized open-source models by over 12% on ROUGE metrics.
Additionally, these models match top closed-source LLMs' performance while being under 1% of their size. We specifically evaluate our framework for oncology, with the potential for adaptation to other specialties.

SpecialtyScribe: Enhancing SOAP note Scribing for Medical Specialties using LLM's

workshop paper

### Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Welcome to the 2025 meeting of the Nations of the Americas Chapter of the Association for Computational Linguistics! I am proud to help organize the first NAACL conference to carry the new name of our organization, one that emphasizes inclusion for all of the Americas. I am also pleased to welcome you to Albuquerque, New Mexico, a state whose unique blend of cultural influences will make for an excellent backdrop for NAACL 2025, especially with this year’s special theme on NLP in a Multicultural World. 
**[Continue reading...](https://drive.google.com/file/d/1jX-qGhqVSZZCIrAnJaz798pflu5Irrdn/view?usp=sharing)**

*- Colin Cherry, Google, NAACL 2025 General Chair* 

[![](https://assets.underline.io/markdown_image/1/image/b087f8a4dc5816d6e1a6514e59c59ac3.png)](https://drive.google.com/file/d/1T96GzPqObXrMln2BMByCSXTSizjTg69P/view?usp=sharing)

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

To access NAACL 2025 event page you are required to register. Please follow [**this link**](https://2025.naacl.org/registration/registration/) to register. Access will depend on your registration type.

Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Reinforcement learning (RL) is increasingly used in the healthcare domain, particularly for the development of personalized adaptive health interventions. However, RL methods are often applied to this domain using small state spaces to mitigate data scarcity. In this paper, we aim to use Large Language Models (LLMs) to incorporate text-based user preferences and constraints, to update the RL policy. The LLM acts as a filter in the action selection. To evaluate our method, we develop a novel simulation environment that generates text-based user preferences and incorporates corresponding constraints that impact behavioral dynamics. We show that our method can take into account the text-based user preferences, while improving the RL policy, thus improving personalization in adaptive intervention.

Using LLMs to improve RL policies in personalized health adaptive interventions

Integrating external knowledge into large language models has demonstrated potential for performance improvement across a wide range of tasks. This approach is particularly appealing in domain-specific applications, such as in the biomedical field. However, the strategies for effectively presenting external knowledge to these models remain underexplored. This study investigates the impact of different knowledge presentation methods and their influence on model performance. Our results show that inserting knowledge between demonstrations helps the models perform better, and improve smaller LLMs (7B) to perform on par with larger LLMs (175B). Our further investigation indicates that the performance improvement, however, comes more from the effect of additional tokens and positioning than from the relevance of the knowledge.

Leveraging External Knowledge Bases: Analyzing Presentation Methods and Their Impact on Model Performance

The remarkable results shown by medical question-answering systems lead to their adoption in real-life applications. The systems, however, may misinform the users, even when drawing on scientific evidence to ground the results. The quality of the answers may be verified by the users if they analyze the evidence provided by the systems. User interfaces play an important role in engaging the users. While studies of the user interfaces for biomedical literature search and clinical decision support are abundant, little is known about users' interactions with medical question answering systems and the impact of these systems on health-related decisions. In a study
of several different user interface layouts, we found that only a small number of participants
followed the links to verify automatically generated answers, independently of the interface design. The users who followed the links made better health-related decisions.

Will Gen Z users look for evidence to verify QA System-generated answers?

Summarizing clinical trial data poses a significant challenge due to the structured, voluminous, and domain-specific nature of clinical tables. While large language models (LLMs) such as ChatGPT, Llama, and DeepSeek demonstrate potential in table-to-text generation, they struggle with raw clinical tables that exceed context length, leading to incomplete, inconsistent, or imprecise summaries. These challenges stem from the structured nature of clinical tables, complex study designs, and the necessity for precise medical terminology. To address these limitations, we propose an end-to-end pipeline that enhances the summarization process by integrating fact selection, ensuring that only the most relevant data points are extracted for summary generation. Our approach also incorporates a feedback-driven refinement mechanism, allowing for iterative improvements based on domain-specific requirements and external expert input. By systematically filtering critical information and refining outputs, our method enhances the accuracy, completeness, and clinical reliability of generated summaries while reducing irrelevant or misleading content. This pipeline significantly improves the usability of LLM-generated summaries for medical professionals, regulators, and researchers, facilitating more efficient interpretation of clinical trial results. Our findings suggest that targeted preprocessing and iterative refinement strategies within the proposed piepline can mitigate LLM limitations, offering a scalable solution for summarizing complex clinical trial tables.

LLM Based Efficient CSR Summarization using Structured Fact Extraction and Feedback

decision support are abundant, little is known about users' interactions with medical question answering systems and the impact of these systems on health-related decisions. In a study of several different user interface layouts, we found that only a small number of participants followed the links to verify automatically generated answers, independently of the interface design. The users who followed the links made better health-related decisions.

MNLP at PerAnsSumm: A Classifier-Refiner Architecture for Improving the Classification of Consumer Health User Responses

The PerAnsSumm Shared Task at CL4Health@NAACL 2025 focused on Perspective-Aware Summarization of Healthcare Q/A forums, requiring participants to extract and summarize spans based on predefined perspective categories. Our approach leveraged LLM-based zero-shot prompting enhanced by semantically-similar In-Context Learning (ICL) examples. Using Qwen-Turbo with 20 exemplar samples retrieved through NV-Embed-v2 embeddings, we achieved a mean score of 0.58 on Task A (span identification) and Task B (summarization) mean scores of 0.36 in Relevance and 0.28 in Factuality, finishing 12th on the final leaderboard. Notably, our system achieved higher precision in strict matching (0.20) than the top-performing system, demonstrating the effectiveness of our post-processing techniques. In this paper, we detail our ICL approach for adapting Large Language Models to Perspective-Aware Medical Summarization, analyze the improvements across development iterations, and finally discuss both the limitations of the current evaluation framework and future challenges in modeling this task. We release our code for reproducibility.

Lightweight LLM Adaptation for Medical Summarisation: Roux-lette at PerAnsSumm Shared Task

The PerAnsSumm 2024 shared task at the CL4Health workshop focuses on generating structured, perspective-specific summaries to enhance the accessibility of health-related information. Given a Healthcare community QA dataset containing a question, context, and multiple user-answers, the task involves identifying relevant perspective categories, extracting spans from these perspectives, and generating concise summaries for the extracted spans. We fine-tuned open-source models such as Llama-3.2 3B, Llama-3.1 8B, and Gemma-2 9B, while also experimenting with proprietary models including GPT-4o, o1, Gemini-1.5 Pro, and Gemini-2 Flash Experimental using few-shot prompting. Our best-performing approach leveraged an ensemble strategy, combining span outputs from o1 (CoT) and Gemini-2 Flash Experimental. For overlapping perspectives, we prioritized Gemini. The final spans were summarized using Gemini, preserving the higher classification accuracy of o1 while leveraging Gemini's superior span extraction and summarization capabilities. This hybrid method secured fourth place on the final leaderboard among 100 participants and 206 submissions.

AICOE at PerAnsSumm 2025: An Ensemble of Large Language Models for Perspective-Aware Healthcare Answer Summarization

Healthcare community question-answering (CQA) forums have become popular for users seeking medical advice, offering answers that range from personal experiences to factual information. Traditionally, CQA summarization relies on the best-voted answer as a reference summary. However, this approach overlooks the diverse perspectives across multiple responses. Structuring summaries by perspective could better meet users' informational needs. The PerAnsSumm shared task addresses this by identifying and classifying perspective-specific spans (Task_A) and generating perspective-specific summaries from question-answer threads (Task_B). In this paper, we present our work on the PerAnsSumm shared task 2025 at the CL4Health Workshop, NAACL 2025. Our system leverages the RoBERTa-large model for identifying perspective-specific spans and the BART-large model for summarization. We achieved a Macro-F1 score of 0.9 (90%) and a Weighted-F1 score of 0.92 (92%) for classification. For span matching, our strict matching F1 score was 0.21 (21%), while proportional matching reached 0.68 (68%), resulting in an average Task A score of 0.6 (60%). For Task B, we achieved a ROUGE-1 score of 0.4 (40%), ROUGE-2 of 0.18 (18%), and ROUGE-L of 0.36 (36%). Additionally, we obtained a BERTScore of 0.84 (84%), METEOR of 0.37 (37%), and BLEU of 0.13 (13%), resulting in an average Task B score of 0.38 (38%). Combining both tasks, our system achieved an overall average score of 49% and ranked 6th on the official leaderboard for the shared task.

LTRC-IIITH at PerAnsSumm 2025: SpanSense - Perspective-specific span identification and Summarization

This paper presents the approach of the DataHacks team in the PerAnsSumm Shared Task at CL4Health 2025, which focuses on perspective-aware summarization of healthcare community question-answering (CQA) forums. Unlike traditional CQA summarization, which relies on the best-voted answer, this task captures diverse perspectives, including ‘cause,' ‘suggestion,' ‘experience,' ‘question,' and ‘information.' The task is divided into two subtasks: (1) identifying and classifying perspective-specific spans, and (2) generating perspective-specific summaries. We addressed these tasks using Large Language Models (LLM), fine-tuning it with different low-rank adaptation (LoRA) configurations to balance performance and computational efficiency under resource constraints. In addition, we experimented with various prompt strategies and analyzed their impact on performance. Our approach achieved a combined average score of 0.42, demonstrating the effectiveness of fine-tuned LLMs with adaptive LoRA configurations for perspective-aware summarization.

DataHacks at PerAnsSumm 2025: LoRA-Driven Prompt Engineering for Perspective Aware Span Identification and Summarization

In this paper, we describe our submission for the shared task on Perspective-aware Healthcare Answer Summarization. Our system consists of two quantized models of the LlaMA family, applied across fine-tuning and few-shot settings. Additionally, we adopt the SumCoT prompting technique to improve the factual correctness of the generated summaries. We show that SumCoT yields more factually accurate summaries, even though this improvement comes at the expense of lower performance on lexical overlap and semantic similarity metrics such as ROUGE and BERTScore. Our work highlights an important trade-off when evaluating summarization models.

Downloads

Next from NAACL 2025

Using LLMs to improve RL policies in personalized health adaptive interventions

Similar lecture

Exploring Robustness in Doctor-Patient Conversation Summarization: An Analysis of Out-of-Domain SOAP Notes